Retinal imaging study uses machine learning to identify Parkinson’s markers

Researchers at the Moorfields Eye Hospital and the UCL Institute of Ophthalmology have discovered “markers that indicate the presence of Parkinson’s disease in patients on average seven years before clinical presentation”, in “the largest study to date on retinal imaging in Parkinson’s disease”.

In a study published in the Neurology medical journal, researchers share how they identified markers of Parkinson’s in eye scans from the AlzEye dataset, with the help of AI machine learning. They repeated the analysis using the wider UK Biobank database of healthy volunteers, which replicated their results.

Post-mortem examination of patients with Parkinson’s disease had previously found differences in the inner nuclear layer (INL) of the retina, with the authors of the new study sharing that their research confirms “previous reports of a significantly thinner ganglion cell-inner plexiform layer, whilst for the first time finding a thinner INL”. This, they say, indicates that “reduced thickness of these layers was associated with increased risk of developing Parkinson’s disease”.

Commenting the potential impacts of the study in real world settings, Moorfields notes that high-resolution images of the retina are now a routine part of eye care, particularly a type of 3D scan known as optical coherence tomography (OCT) which can “produce a cross-section of the retina in incredible detail – down to a thousandth of a millimetre”. OCT scans are already “widely used in eye clinics and high-street opticians”, and the researchers note that the utilisation of machine learning enables them to “accurately analyse large numbers of OCTs and other eye images, in a fraction of the time it would take a human”.

Siegfried Wagner, clinical research fellow at Moorfields Eye Hospital, and principal investigator, comments: “I continue to be amazed by what we can discover through eye scans. While we are not yet ready to predict whether an individual will develop Parkinson’s, we hope that this method could soon become a pre-screening tool for people at risk of disease.”

Earlier this year, at HTN’s AI and Data event, we were joined by Peter Thomas, chief clinical information officer and director of digital medicine at Moorfields Eye Hospital, for a discussion on building the foundations of scaled AI implementation. Peter discussed the applications of AI in eye care, and considered how the broader healthcare system might have to change in order to ensure that implementations are successful and safe.

Elsewhere in the use of AI for early diagnosis of Parkinson’s, our news in brief from July highlighted work done by researchers from Cardiff University, who discovered that by using artificial intelligence to analyse smart watch data, they could “accurately predict those who would go on to later develop Parkinson’s disease”.